Physical Characterization of the Helium-Oxygen-Acetylene Flame as an Atom Cell for Atomic Spectrometry K. A. Saturday and G. M. Hieftje" Department of Chemistry. Indiana University, Bloomington, Indiana 4 740 1
Phydcal characterizatlon of a heiium-oxygen-acetylene flame supported on a water-cooled burner has shown the flame to possess a high degree of laminarlty and a maximum safe operating temperature of 2760 K (with 60% helium diiuent). Safe operating conditions for the flame have been established to enable its safe and flashback-free operation. I n addltlon, the spectral background of the flame has been examined and found to be relatively low compared to other high temperature flames, suggesting its use as a promislng alternative atom cell for atomic emission, absorption, or fluorescence spectrometry.
Functioning as a combined atom source, sample cell, and excitation medium for atomic spectrometry, the typical analytical flame forms an environment which has a complex and dynamic, yet easily utilizable character unparalleled by any other analytical source. In this application, the flame serves a number of functions essential to flame spectrometric analysis, including desolvation of the sample aerosol, vaporization of solute particles, and atomization of the vaporization products. Unfortunately, flames are notoriously inefficient and poorly understood in how they accomplish these results, and only through a basic understanding of how each step is accomplished will the flame's utility be improved markedly over that now attainable. In the past, improvements in flames were derived largely through empirical means. Variations in flame-gas constituents and their proportions were found to be most effective in improving the flame, although burner design, the use of sheathing gases, and the development of different flame shapes were also found to be important. However, because of the complex environment within each flame and the interrelationships of variables affecting that environment, it was difficult or impossible to single out individual flame properties for optimization. Clearly, what would be desirable would be to separate all variables, characterize them, and employ the results of such characterization to generate improvements. In our laboratory, just such investigations have been under way for some time, and the present report indicates the utility of those findings in developing new and improved flames for use in spectrometric analysis. In recent studies (1-3), we investigated the processes of droplet desolvation and solute particle vaporization in an analytical flame. These studies have yielded quantitative relationships describing those processes and have increased our knowledge of atom formation to the point where some meaningful changes could be made to increase the flame's performance. In a study on the mechanism of droplet desolvation in a flame (I,3), it was found that flame-gas thermal conductivity played an important role in the desolvation process. A subsequent study demonstrated that replacement of the air component in an air-acetylene flame with a high thermal-conductivity oxidizer gas mixture (helium + oxygen) increased the aerosol desolvation and sample utilization efficiencies (3). Unfortunately, with these advantages came an increased occurrence of flashback, thermal deformation of the
burner components, and the high cost of the helium diluent gas ( 3 , 4 ) . Besides its increased sample utilization and atom formation efficiencies, the helium-oxygen-acetylene flame holds the possibility of increased fluorescence yield because of its low concentration of quenching species. Unlike most analytical flames, the helium-xygen-acetylene flame has as its primary constituent a monatomic gas (helium),which has an unusually low collisional cross section (5). Helium-diluted hydrogenoxygen flames have already been shown to exhibit higher quantum efficiencies than similar nitrogen-diluted flames (6); contrasted to these previously studied flames, the heliumoxygen-acetylene flame should produce higher atom concentrations, because of the large number of electrons and carbon-containing radicals it produces (7). T o offset this advantage somewhat, the acetylene-fueled flame would be expected to exhibit greater quenching than a similar hydrogen-fueled flame. The present study seeks to increase the utility of the helium-oxygen-acetylene flame by developing a burner system to produce a stable, reproducible flame and to establish safe operating conditions for the new system. Physical characteristics of the resulting flame system were investigated in order to compare the new flame with others commonly employed. Laminarity of the flame was investigated using Schlieren photography, and temperature profiles of an analytically useful flame were obtained. For practical applications, the background spectrum of the new flame was compared to that of a commonly used air-acetylene flame.
EXPERIMENTAL The burner and gas flow system used with the He-02-C2H2 flame are described in this section. In addition, special techniques used to characterize the flame are discussed. Burner. A modified version of the Meker-type burner described by Hieftje (8) was used to support the new flame. The 1-cm thick burner top shown in Figure 1 was constructed of nickel-plated brass and contained 169 exit ports of 0.45-mm diameter, which were arranged in concentric rings. Port diameter was chosen on the basis of a trade-off between safety and convenience in operation and fabrication; the 0.45-mm diameter ports are small enough to prevent flashback when used under the safe operating conditions defined in this paper, but large enough for easy machinability and for avoidance of clogging by combustion products and sample residue. Silver-solderedto the burner top was a single turn of copper tubing (- 5-mm id.) through which tap water was passed at 0.5 L/min. At this flow, the cooling coil absorbed sufficient heat from the burner top to eliminate thermal deformation of the burner and to further reduce the likelihood of flashback. The operating temperature of the water-cooled burner, which was found to depend upon total gas flow and flame stoichiometry, reached a maximum of 650 K for a flame near flashback. The typical equilibrium temperature of the burner used with a flame produced under safe operating conditions was 425 K. Flame Gases. As shown in Figure 2, purified helium, oxygen (Matheson Co., Joilet, Ill.) and acetylene (Linde Division, Union Carbide Corp., New York, N.Y.) were supplied through single-stage regulators (Matheson Co.) and measured with dual-float rotameters (He: Matheson Co.; O2and C2Hz:Dwyer Manufacturing ANALYTICAL CHEMISTRY, VOL. 49, NO. 13, NOVEMBER 1977
2013
BURNER
BOWL
BURNER T C P
0 4 5 mm DIAM
COOLING WATER
COPPER
(Jmm
I.--35Crn--* '4
TUBING d)
,
--48cm-
Figure 1. Top view of water-cooled burner used for the heliumoxygen-acetylene flame. A cooling coil was silver-soldered onto a burner to which a detachable burner head containing 169 0.45mmdiameter ports was fastened
Figure 2. Gas-flow control system used to support the heliumoxygen-acetylene flame showing arrangement for mixing the helium and oxygen gas flows
Inc., Michigan City, Ind.). The calibration of each flowmeter was established by a wet-test meter (Precision Scientific Co.; Chicago, Ill.). Needle valves and toggle shut-off valves (Nupro Co., Cleveland, Ohio) were used to control the gas flows. After exiting from their flowmeters, the helium and oxygen were mixed at a tubing '3"' and fed through a coiled tube ( 4 ) to ensure good mixing before the gases entered the burner chamber. The acetylene was fed directly to the burner from the flow system. When an airacetylene flame was employed, the air (Matheson Co.) was also fed directly to the burner from the rotameter (Dwyer Manufacturing Inc.). All flows reported in this paper have been corrected to STP. Determination of Safe Operating Conditions. The limits of safe operation of the new burner and flame were ascertained by determining the flame gas flows necessary to prevent flashback. To perform these measurements, a siable flame was established which met the desired requirements of fuel-to-oxidant flow ratio or oxygen flow rate Small changes in the helium flow were then made until flashback occurred. After each decrement in helium flow, thermal equilibrium was allowed to establish at the burner top before further flow adjustments were made. Thermal equilibrium was judged on the basis of a stable reading from a chromel-alumel thermocouple attached to the burner head. After each flashback was incurred, flow readings were taken. Flashback flow measurements were taken several times for each set of flame conditions to ensure reproducibility of the data. Schlieren Photography. Schlieren photographs (9) were produced using two spherical mirrors designed for that purpose and obtained from Edmund Scientific Co., Barrington, N.J. A pin-hole light source for the photographs was constructed by focusing the tungsten-halogen lamp from a 35-mm slide projector (model AF Carousel, Kodak) onto an iris diaphragm which was placed at the focus of the first spherical Schlieren mirror. The collimated beam produced by the first mirror was passed through the flame to a second spherical mirror which focused the light at a circular aperture (IO). Only those light rays which remained unrefracted as they passed through the flame were allowed to pass through the circular aperture and to the focal plane of a photographic film holder placed behind the aperture. A 1-msexposure time was used for each photograph in order to allow viewing of the greatest possible portion of the turbulence which would affect a typical analytical determination. Further detaibabout the Schlieren technique can be found in the literature (9, IO). Flame Temperatures. Flame temperatures were measured using the sodium line-reversal technique (11). In these measurements, a tungsten strip lamp served as the primary source and was calibrated with an optical pyrometer (Pyrometer In2014
ANALYTICAL CHEMISTRY, VOL. 49, NO. 13, NOVEMBER 1977
strument Co., Northvale, N.J.) whose validity is traceable to the National Bureau of Standards. To further assure accuracy, the pyrometer was periodically checked with an NBS-traceable tungsten-filament lamp (No. 18AT10/2, General Electric Co., Cleveland, Ohio). Detection of the reversed condition (indicating equality of the lamp and flame temperatures), as described by Snelleman ( I I ) , was simplified by use of a wavelength-modulated monochromator and synchronous detection system (12). R E S U L T S AND DISCUSSION The D e t e r m i n a t i o n of S a f e O p e r a t i n g Limits: Flashback Studies. Results of an earlier study ( 3 ) have shown the He-02-CzH2 flame t o be flashback-prone when supported on a burner designed for operation with an airacetylene mixture. This finding is not surprising if one considers the factors affecting the formation of a pre-mixed, laminar flame. The stabilization of such a flame on a burner top is governed by three processes: first, the flame gases are consumed a t a rate (characterized by their burning uelocity) which is directly related to the composition of the flame gas mixture; second, the walls of the burner ports exert a quenching effect on the combustion of the gas mixture, making the burning velocity zero near the port walls; third, a short distance from the port walls, the flame gases again attain their characteristic burning velocity (9, 13). For a gas exhibiting laminar flow out of a burrier port, a Poiseuille velocity distribution exists across the port diameter. T o prevent flashback, the exiting gas velocity at every point in the burner port must be greater than or equal to the backward-propagating burning velocity at that point. The gas flow which produces a velocity profile which is everywhere just equal to or greater than the burning velocity is said t o generate the critical boundary velocity gradient. This flow is then the lowest which will prevent flashback. Flashback occurs when the gases flowing from the burner port do not exceed the critical boundary velocity gradient, causing the flame to propagate into the burner's mixing chamber and detonate the gas mixture there. Obviously, the larger the burner port or the greater the burning velocity of the gas mixture, the greater will be the gas velocity or flow required to prevent flashback. Because the He-02-C2H2 mixture has a far greater burning velocity than air-acetylene, i t is not surprising that the former flame could not be safely supported on a burner designed for air-acetylene. The general approach to producing a safe and flashback-free flame can take two courses: the burner ports can be made small enough to extend the quenching effects of the walls throughout the port, or the minimum gas velocity required t o maintain a safe flame can be determined for a particular burner port size. The first of these courses is often not practicable for flames of the kind being studied here. In order t o calculate the port diameter which is small enough to maintain the flame under the worst, possible conditions, a knowledge of the combustion characteristics for all possible combinations of gas mixture must be available. When a burner port exhibits this behavior, it is said to be a t or below the quenching diameter for t h e gas mixture(a) used. Quenching diameter values for several flames have been calculated or empirically determined (8, 14, 15). Unfortunately, for rapidly burning gas mixtures, the required quenching diameters are often very small. For example, an oxygen-acetylene mixture has a quenching diameter of 0.406 mm (15). Such small ports are generally impractical because of the difficulty of machining a multiport burner head and the likelihood of clogging of the ports with nebulized solution and combustion products. In this situation, the second approach must be taken and safe operating conditions for the flame must be empirically determined. This latter method is the one ordinarily adopted by instrument manufacturers who offer burners for nitrous oxide-acetylene flames.
SOD c L A S H BACK
120-
PREVENTED
5
400
FLASHBACK
t
OCCbRS
4
' 330 1 30 0 1 02 03
04
05
OS
\ O B
07
l o PREVENTED
2 OS
1 f
1: 75
FLASHBACK
OCCURS
RAT 0
FUELlOXlDANT
iro
80
t
Variation in the helium-requirement as a function of fuel/ oxidant ratio for a helium-oxygen-acetylene flame. The minimum helium requirement to maintain a flame above flashback is noted for a flame gas mixture containing 4.1 L/min oxygen. FueVoxidant variations were accomplished by changing the acetylene flow Flgure 3.
6:o
-
f
*
-
50
-
.d
4'0
Flgure 5. Effect of oxygen flow on t h e total flow requirement for a helium-oxygen-acetylene flame with a fueVoxidant ratio of 0.49 I30
55c
2
-
r
-I
0
7 5
PREYLliTED
PREVEVTEC
5
0
5CO-
I10
c
i6 5
D '\
45c
-
t
I
Z _I
I
9 o c
C
-
4 0 C - - L 00
0
1
1
30
40
20 O X l G E Y CLOW
1 5 0
''
1' 60
5
FLASHBACK OCCURS
I
' i
/mn1
Flgure 4. Effect of oxygen flow on the hellum requirement of the hellum-oxygen-acetylene flame. A constant fuel/oxldant ratio of 0.49 was maintained for all oxygen flows
The critical boundary velocity gradient for differing flame compositions depends not only on the burning velocity of the flame gases but also on their density, thermal conductivity, and specific heat (16). For a flame composed of gases whose composition is either pure (Le., CzH2)or constant (Le., air), the burning velocity has been shown to change relatively little over a wide range of fuel/oxidant ratios, when compared with the large changes in burning velocity which occur when varying amounts of a diluent gas are added to the same gas mixture (9, 17). Therefore, varying the percentage of diluent gas in a flame gas mixture is one method of controlling the burning velocity of the flame. For instance, an acetylene flame supported by a 21% 0 2 mixture with N2 has been shown to have a burning velocity one-tenth that of a flame supported by pure O2(17). A similar study substituted helium for the nitrogen diluent and found the helium-diluted flame to have a burning velocity over twice that of the nitrogen diluted flame (18). For all mixtures, i t was found that the burning velocity decreased as the fraction of diluent increased. Since the He-OZ-C2Hz flame operates with variable helium flow, the maximum burning velocity of the flame is that of a pure oxyacetylene flame (1130 cm/s) (17). The helium diluent gas serves not only to increase the desolvation capabilities of the flame but also to lower the burning velocity to the point where the flame can be more easily stabilized from flashback. The amount of helium required in the present study to maintain a stable flame for various acetylene/oxygen flow ratios (Le., fuel/oxidant ratios) is shown in Figure 3. T o obtain the data in Figure 3, the oxygen flow was maintained
7.0
L
60 1 40
4 I
45
4o
1 50
55
PER CENT HELilJM
Flgure 6. Safe operating condttions for the helium-oxygen-acetylene flame. A constant fuel/oxidant ratio of 0.49, which represents a worstcase situation, was used to delineate the safe operating conditions for all useful flame gas flows
a t 4.1 L/min; the variation in fuel/oxidant was accomplished by changing the acetylene flow rate.. The slight changes in burning velocity of the mixture so produced were countered by altering the amount of helium diluent gas required to prevent flashback. Interestingly, the total gas flow required to prevent flashback was found to remain nearly constant a t 11.5 L/min between the fuel/oxidant ratios of 0.45 to 0.87, with slightly lower total flows required at the low fuel/oxidant ratios. The data displayed in Figure 3 follow a predictable pattern. At the gas-flow ratio characteristic of the stoichiometric production of COz (fuel/oxidant = 0.40), the greatest percentage of the oxygen and acetylene components react in the combustion process, to produce a maximal burning velocity. However, air entrainment, burner top heating, and incomplete combustion result in a shifting of this maximum burning velocity point to slightly richer measured gas mixtures. At fuel/oxidant ratios away from stoichiometric, the gas burning velocity decreases, as signalled by the smaller percentage of helium required to prevent flashback. This behavior suggests that portions of the acetylene or oxygen flows are not actively ANALYTICAL CHEMISTRY, VOL. 49,NO. 13, NOVEMBER 1977
2015
Schlieren photographs of (a) He-02-C2H2flame and (b) Bunsen flame. The dark regions denote refractive Index changes which can occur as a result of turbulent mixing Flgure 7.
involved in the combustion process and act as diluent gases. Figure 3 reveals that the helium content of the H e 0 2 - C 2 H 2 flame gas mixture must be nearly 50% to prevent flashback a t all values of fuel/oxidant ratio. In addition, if the oxygen flow is changed, the helium requirement is also altered as shown in Figure 4. T o construct Figure 4, a constant fuel/oxidant ratio of 0.49 was chosen to represent the region of greatest helium requirement (cf. Figure 3); also, oxygen and acetylene flows were reduced simultaneously in order to maintain a fixed fuel/oxidant ratio. Thus, the increase in oxygen flow (horizontal axis of Figure 4) also represents an increase in the total flow of active components in the gas mixture. From Figure 4, lower oxygen and acetylene flow rates require correspondingly more helium to prevent flashback. It might then be expected that a constant total flow (including helium) for all oxygen/acetylene combinations would be required. However, Figure 5 shows that the total required flow decreases monotonically with oxygen flow; obviously, the increased dilution of the combustible gases by helium reduces the mixture's burning velocity. Useful operating limits of the helium-oxygen-acetylene flame are delineated in Figure 6. Because the fuel/oxidant ratio of 0.49 represents the greatest demand for helium (cf. Figure 3), the percent helium required for a specific total flow a t this fuel/oxidant ratio will suffice for any other fuel/oxidant ratio. Of course, other fuel/oxidant ratios will permit the use of flame gas mixtures with lower helium content. The lowest helium demand is at the extremes of the useful fuel/oxidant range which extends between the point at which the flame turns luminous (fuel/oxidant = 0.97) and where the acetylene flow becomes too low to be accurately monitored (fuel/oxidant = 0.25). At these fuel/oxidant ratios, the helium demand is 28% and 45%, respectively. However, as a general rule, it can be ascertained from Figure 6 that a typical analytical flame (7 L/min total flow or more), supported on a burner head with 0.45-mm diameter ports, will operate safely with a mixture of helium, oxygen, and acetylene containing 55% or more helium. F l a m e Laminarity: Schlieren Studies. Schlieren methods have long been used to observe spatial refractive index changes within flames and other combustion sources (9,10). Information obtained from Schlieren photographs has been used to correlate turbulence in the flame background radiation to noise seen in an atomic absorption measurement 2016
ANALYTICAL CHEMISTRY, VOL. 49, NO. 13, NOVEMBER 1977
2800 r
2700
1
A
I
,"
2500 r
;2 6 0 0
D
p 2400
-
2300
k
7
,
2200
030
040
010
060
070
FUE./OXIDANT
080
090
I00
RAT10
Flgure 8. Variation of helium-oxygen-acetylene flame temperature with fueVoxidant ratlo at (A) 1 cm above the burner top and (B) 8 cm above the burner top. A constant oxygen flow of 4.1 L/min and a
constant total flow of 19.6 L/min were maintained
(19). In the present study, the laminarity of the heliumoxygen-acetylene flame was investigated using Schlieren methods. The Schlieren image produced by the helium-oxygenacetylene flame (Figure 7A) is identical to that of an airacetylene flame which was found by others to exhibit a high degree of laminarity (8, 11, 20). A slight turbulence at the edges of the flame and in its upper regions observable in Figure 7A is caused by the entrainment of air. However, the central portion of the flame remains fairly laminar throughout the most commonly used region (within 10 cm above the burner top). For comparison, the Schlieren image of a Bunsen flame is shown in Figure 7B. The regions of turbulence within the Bunsen flame produced by the mixing of flame gases and air are easily observed in the high-contrast regions of the image (9). The wide degree of refractive index change in the mixing gases produces the sharp contrasts. F l a m e T e m p e r a t u r e s . The importance of flame temperature to practical flame spectrometry has long been known. The use of the nitrous oxide-acetylene flame illustrates the increase in sensitivity and decrease in matrix interferences that higher flame temperatures can bring, but also betrays the effects of increased ionization (21). Recent studies have shown the important role flame temperature plays in the desolvation and vaporization processes in flames (1-3).
Ha-02-C2H2
FLAME
x5
I li 3000
Air-C2H2
35 00
4000
FLAME
5000
4500
WAVELENGTH
5500
6000
(AI
Figure 9. Background spectra of air-acetylene and helium-oxygen-acetylene flames of equal fuel-oxidant ratio
The change in He-02-C2H2 flame temperature with fuel/oxidant ratio a t two different heights in the flame is shown in Figure 8. To produce Figure 8, the total gas and oxygen flow rates were kept constant, but the acetylene and helium flow rates were varied oppositely to produce the desired fuel/oxidant ratio and total flow. T h e results show a rise in temperature with increasingly richer flames and decreasing helium content. From Figure 8, the flame temperatures exhibited by the He-02-CzH2 flame are higher than those of an air-acetylene flame and approach those of a nitrous oxideacetylene flame, which reach maximum flame temperatures of 2550 K (C/O = 0.50) (11) and 3070 K (C/O = 0.09) (221, respectively. Addition of a shielding flame should enhance the temperature of the He-02-C2H2 flame by reducing the rate of heat loss just as has been found for other flames (11). The flame also offers the ability to vary the temperature over a 200-300 K range by careful choice of fuel/oxidant ratio and helium content. This ability enables an appropriate temperature and flame composition to be selected independently for the most sensitive determination of any element. The vertical temperature profile of the He-02-C2H2flame was investigated at a flame composition corresponding to the highest flame temperature displayed in Figure 8 (i.e., fuel/ oxidant = 0.85). The temperature was found to decrease linearly with a gradient of approximately 34 K per centimeter from the 2760 K temperature measured 1cm above the burner head. Background Spectra. A comparison between the background spectrum of the helium-oxygen-acetylene flame and that of an air-acetylene flame with equal acetylene/ oxygen ratio and total flow rate reveals the similarity of the two flames. Using identical conditions of spectral bandpass (0.38 mm) and detector sensitivity, the spectra shown in Figwe 9 were obtained just above the primary reaction zone. Both flames exhibit the characteristic OH band a t 3070 A, Swan C2 bands a t 4680 A and 5160 A, and CH bands a t 3870 A and 4310 A (23).Although both spectra are quite simple compared to that of a nitrous oxide-acetylene flame, the background of the helium-oxygen-acetylene flame is about four times as intense as that of the air-acetylene flame a t the strongest spectral feature observed (the OH band a t 310 nm). However, this intensity ratio would be expected to be much greater if only the temperature difference between the two flames were taken into consideration. This disparity indicates that a gain
in elemental emission signal/ background intensity should be observed for the He-O2-C2H2 flame over the air-C2H2 flame. Studies are currently under way to substantiate this hypothesis.
CONCLUSIONS With the development of a water-cooled burner, the helium-oxygen-acetylene flame has been characterized to allow its safe and flashback-free use in atomic spectrometry. The flame offers physical characteristics that are similar to or better than the &acetylene flame in the areas of temperature and laminarity while yielding a slight disadvantage in background emission intensity. It should offer increased sensitivity for the determination of refractory elements while not possessing the high background and incidence of ionization which plague the nitrous oxide-acetylene flame. The incorporation of helium into the oxygen-acetylene flame gas mixture has added a new control dimension to the flame. The third variable gas flow enables the independent control of such flame properties as burning velocity, temperature, and rise velocity a t any selected fuel/oxidant ratio. The Came has a wide range of fuel/oxidant ratios available which should allow optimization of flame conditions for many elements. In use, the operating costs of the helium-containing flame exceed those of the air-acetylene flame, but are comparable to the costs of a nitrous oxide-acetylene flame. The analytical utility of the helium--oxygen-acetylene flame will be investigated by the measurement of atomization efficiencies, a project which is currently in progress in our laboratories. T h e utility of the flame as an atom source for atomic fluorescence will also be studied by means of fluorescence quantum efficiency measurements; we expect the measurements to substantiate the importance of the flame's monatomic primary constituent to improved fluorescence yields. Future work will also include the development of a flashback-free, air-cooled slot burner to support a long-path laminar He-O2-CzH2 flame for use in atomic absorption spectrometry.
ACKNOWLEDGMENT The authors acknowledge N. C. Clampitt for helpful advice concerning the helium-oxygen-acetylene flame.
LITERATURE CITED (1) N. C. Clampitt and G. M. Hieftje, Anal. Chem., 44, 1211 (1972) (2) G. J. Bastlaans and G. M. Hleftje, Anal. Chem., 48, 901 (1974). ANALYTICAL CHEMISTRY, VOL. 49, NO. 13, NOVEMBER 1977
e
2017
(3) N. C. Clampltt and G. M. Hieflje, A n d . Chem., 46, 382 (1974). (4) N. C. Clampttl, Ph.D. Thesis, Indiana Universky, Blooniington, Ind., 1974. (5) D. R. Jenkins, R o c . R . Soc. London, Ser. A , 306, 413 (1968). (6) D. R. Jenkins, Spectrochlm. Acta, 258, 47 (1970). (7) J. D. Wlnefordner, V. Svoboda, and L. J. Cline, Crn. Rev. Anal. Chem., 1, 233 (1970). (8) G. M. Hieftje, Appl. Spectrosc., 2 5 653 (1971). (9) A. G. Gaydon and H. G. Wolfhard, Flames”, Chapman and Hall, Ltd., London, 1970. (10) F. J. Weinberg, “Optics in Flames”, Butterworth and Co. Ltd., Washington, D.C., 1963. (1 1) W. Snelleman, P h D Thesis, University of Utrecht, 1965, (12) G. M. Hieftje and R. J. Sydor, Appl. Spectrosc., 26, 624 (1972). (13) R. Lewis and G. Von Elbe, ‘Combustion, Flames and Explosions of Geses”, Academlc Press, New York, N.Y., 1951. (14) K. M. Aldous, B. W. Bailey, and J. M. Rankin, Anal. Chem., 44, 191 (1972). (15) R . F. Suddendorf and M. B. Denton, Appl. Spectrosc., 2 8 , 8 (1974). (16) S.B. Reed, Combust. Name, 13, 583 (1969). (17) E. Bartholm6, Z . Electrochem., 54, 169 (1950).
(18) P. J. Wheatley and J. W. Llnnett, Trans. Faraday Soc., 48, 338 (1952). (19) K. M. Aldous, R. F. Browner, D. Clark, R. M. Dagnall, and T. S.West, Talanta, I S , 927 (1972). (20) C. Th. J. Alkemade, Ph.D. Thesis, University of Utrecht, 1954. (21) M. D. Amos and J. B. Willis, Spectrochlm. Acta, 2 2 , 1325 (1966). (22) G. F. Klrkbrlght, M. K. Peters, M. Sargent, and T. S. West, Telenta, 15, 663 (1968). (23) A. G. Oaydon, “The Spectroscopy of Flames”, Chapman and Hall, London, 1974.
RECEIVED for review April 18,1977. Accepted August 5,1977. Presented in part at the second and third national meetings of the Federation of Analytical Chemistry and Spectroscopy Societies. The authors acknowledge the National Science Foundation and the Office of Naval Research for support of this work.
Atomization from a Platform in Graphite Furnace Atomic Absorption Spectrometry D. C. Grigolre and C. L. Chakrabarti’ Metal Ions Group, Department of Chemistry, Carleton University, Ottawa, Ontario, Canada K I S 5B6
Studies on the atomlzatlon of Mol V, Cu, Zn, and Cd from a rectangular pyrolytlc graphlte platform Inserted Into an HGA-2100 pyrolytlcallytoated graphtte tube Is reported. No Improvement In peak-helght sensltlvttles Is observed wtth the exceptlon of Mo and V, for whtch up to 1.8-fold enhancement in the sensltlvlty Is observed. Integrated sensltlvltles are generally enhanced by a factor of 1.2 to 1.4. The effect of platform mass and geometry on the temperature-time characterlstlcs of the platform has been studled. Increaslng the mass of the platform slgnlflcantly decreases the rate of heatlng of the platform at the appearance tlme of any glven metal. Peak absorbance for copper Is severely dlmlnlshed as the platform mass Is Increased. Integrated absorbances, however, are virtually unaffected by even a 5-fold Increase In the platform mass.
Studies done with commercially available atomizers, Varian Techtron Carbon Rod Atomizer 63 (CRA-63) and the Perkin-Elmer Heated Graphite Atomizer 2100 (HGA-2100),show that the conditions necessary for the theoretically predicted maximum attainable sensitivity are not fulfilled (1-3). The basic limiting factors include the geometry of these atomizers as well as their heating characteristics. Transient absorption pulses (2,3) are produced by rapid heating of electrothermal atomizers. The absorbance values can be measured by two methods-the peak absorbance method and the integrated absorbance method ( 4 ) . L’vov ( 4 ) has shown that the peak-height sensitivity is proportional to , the residence time, the ratio of the atomization time, T ~ and 72, of the analyte atoms in the analysis volume. If the atomization is completed in a time much smaller than the residence time (i.e., 7 1 / 7 2
